[unreadable] The European colonial period beginning in the late 1400s brought previously isolated populations together in the New World; thus, many US residents can trace their ancestry to more than one continent. Studying the impact of genetic admixture is important because variation in individual admixture proportions can be a confounding factor in ordinary genetic association studies, and because the genetic structure of admixed populations can facilitate the detection and localization of genes underlying ethnic variation in disease risk. We have demonstrated the usefulness of study designs that exploit the ability to model admixture for: (1) Distinguishing genetic and environmental explanations for ethnic differences in disease risk (and investigating the mode of inheritance), by studying the relationship of disease risk to individual admixture; (2) Controlling for confounding of genetic associations in stratified populations by modeling individual and population admixture; and (3) Admixture mapping: localizing genes that underlie ethnic differences in disease risk. Indeed, one of the greatest areas of promise in admixture studies is gene mapping, which can be accomplished by testing for association between the disease and locus ancestry after conditioning on parental admixture. The advantage of admixture mapping is it has higher statistical power compared to family linkage studies. Application of these approaches to admixed populations such as African Americans, Mexican Americans and Puerto Ricans, not only requires further development of the statistical methods, but also additional markers informative for ancestry. A genome spanning panel of Ancestry Informative Markers (AIMs) is needed for distinguishing European, West African and Native American ancestry. AIMs have the primary feature of large differences in allele frequency between the parental populations. Of the three study designs listed above, typically 100 AIMs from across the genome are needed to estimate individual admixture proportions for (1) and (2). However, design (3) necessitates markers at much higher density because it depends upon inferring ancestry at each locus. From simulation studies, we estimate that about 1000 AIMs will be required for genome-wide mapping with two-way admixture, and about 1500 for admixture between three populations. [unreadable] [unreadable] [unreadable]